FitMyLLM
AMD· RDNA 2.0

AMD Xbox Series X 6nm GPU

Running LLMs on the Xbox Series X 6nm GPU — the long read: which models fit at which quantisation, and the settings worth changing. · Or what a budget buys

VRAM
10 GB
Entry-level
Bandwidth
560
GB/s
Models Q4
238/498
48%
7B Q4 speed
~71
Blazing
Model coverage at Q448% of all
Estimated speed· By model size at Q4

Average speeds at Q4 quantization. Actual performance varies by model architecture and context length.

3B
~166
tok/s
7B
~71
tok/s
14B
~36
tok/s
32B
—
18.0GB needed
70B
—
39.4GB needed
Measured rig reports

We rent the machine and time every model on it: decode, VRAM peak, concurrency, watts and cost per million tokens. Yours may already be one of them — and two are free to read in full.

See the reports →
Specifications
VRAM
10 GB
Bandwidth
560 GB/s
FP16 Compute
24.3 TFLOPS
TDP
200W
Memory
GDDR6
Architecture
RDNA 2.0
Stream Processors
3,328
Compute Units
52
238
Fast models · >30 tok/s
Real-time chat speed
238
Usable · >10 tok/s
Comfortable for all tasks
238
Total compatible
Fit in VRAM at Q4
Don’t want to buy?

Test Xbox Series X 6nm GPU (or anything bigger) without committing. Pay by the second, cancel anytime.

Spin up in ~60s. Pay by the second. Cancel anytime.

Some links are affiliate links — we may earn a small commission at no extra cost to you. This helps keep FitMyLLM free and independent.

Compatible models· 238
S
Falcon-H1R Tiny 90M0.09B
Falcon·256K ctx· chat· reasoning
5531
tok/s · 5% VRAM
›
S
SmolLM2 135M0.135B
Smollm·2K ctx· chat
3687
tok/s · 6% VRAM
›
S
nomic-embed-text-v1.5 100M0.14B
Embedding·8K ctx· chat
3556
tok/s · 6% VRAM
›
S
GPT-2 124M0.14B
Gpt2·1K ctx· chat
3556
tok/s · 6% VRAM
›
S
SmolVLM 256M0.256B
Smollm·8K ctx· chat· vision
1944
tok/s · 6% VRAM
›
S
Gemma 3 270M0.27B
Gemma·32K ctx· chat
1844
tok/s · 7% VRAM
›
S
Snowflake Arctic Embed M v2.00.305B
Embedding·8K ctx· embedding· multilingual
1632
tok/s · 7% VRAM
›
S
bge-large-en-v1.5 335M0.335B
Embedding·1K ctx· chat
1486
tok/s · 7% VRAM
›
S
mxbai-embed-large-v10.335B
Embedding·1K ctx· embedding
1486
tok/s · 7% VRAM
›
S
Snowflake Arctic Embed L0.335B
Embedding·1K ctx· embedding
1486
tok/s · 7% VRAM
›
S
LFM2 350M0.35B
Lfm·125K ctx· chat· tool_use
1422
tok/s · 7% VRAM
›
S
SmolLM2 360M0.36B
Smollm·8K ctx· chat
1383
tok/s · 7% VRAM
›
S
GPT-2 Medium 345M0.38B
Gpt2·1K ctx· chat
1310
tok/s · 7% VRAM
›
S
Qwen 2.5 0.5B0.5B
Qwen·32K ctx· chat
996
tok/s · 8% VRAM
›
S
SmolVLM 500M0.5B
Smollm·8K ctx· chat· vision
996
tok/s · 8% VRAM
›
S
Falcon-H1 0.5B0.52B
Falcon·128K ctx· chat
957
tok/s · 8% VRAM
›
S
BGE-M30.568B
Embedding·8K ctx· embedding
876
tok/s · 8% VRAM
›
S
Snowflake Arctic Embed L v2.00.568B
Embedding·8K ctx· embedding· multilingual
876
tok/s · 8% VRAM
›
S
Qwen3-Embedding 0.6B0.6B
Embedding·32K ctx· embedding· multilingual
830
tok/s · 9% VRAM
›
S
Falcon-H1R Tiny 0.6B0.6B
Falcon·32K ctx· chat· reasoning
830
tok/s · 9% VRAM
›
S
Falcon Perception 0.6B0.6B
Falcon·4K ctx· vision
830
tok/s · 9% VRAM
›
S
Qwen 1.5 0.5B0.62B
Qwen·32K ctx· chat
803
tok/s · 9% VRAM
›
S
LFM2 700M0.74B
Lfm·125K ctx· chat· tool_use
673
tok/s · 9% VRAM
›
S
Qwen3 0.6B0.75B
Qwen·32K ctx· chat· reasoning
664
tok/s · 9% VRAM
›
S
GPT-2 Large 774M0.81B
Gpt2·1K ctx· chat
615
tok/s · 10% VRAM
›
S
Qwen 3.5 0.8B0.87B
Qwen·256K ctx· chat· coding· multilingual
572
tok/s · 10% VRAM
›
S
InternVL3 1B0.94B
Other·32K ctx· chat· vision
530
tok/s · 11% VRAM
›
S
MiniCPM5 1B1.08B
Minicpm·128K ctx· chat· reasoning· multilingual
461
tok/s · 11% VRAM
›
S
K2 Horizon 0.9B1.08B
Other·128K ctx· chat· coding· reasoning
461
tok/s · 11% VRAM
›
S
TinyLlama 1.1B1.1B
Llama·2K ctx· chat
453
tok/s · 12% VRAM
›
S
LFM2.5-1.2B-Thinking1.2B
Lfm·122K ctx· chat· reasoning· tool_use
415
tok/s · 12% VRAM
›
S
Llama-3.2-1B1.2B
Llama·4K ctx· chat
415
tok/s · 12% VRAM
›
S
LFM2 1.2B1.2B
Lfm·125K ctx· chat· tool_use· multilingual
415
tok/s · 12% VRAM
›
S
Zamba2 1.2B1.2B
Other·4K ctx· chat
415
tok/s · 12% VRAM
›
S
Ling 3.0 Tiny7.89BMoE
Other·128K ctx· chat· coding· reasoning
383
tok/s · 53% VRAM
›
S
EXAONE-4.0-1.2B1.3B
Exaone·64K ctx· chat
383
tok/s · 13% VRAM
›
S
OPT 1.3B1.3B
Opt·2K ctx· chat
383
tok/s · 13% VRAM
›
S
MiniCPM-V 4.61.3B
Other·256K ctx· chat· vision
383
tok/s · 13% VRAM
›
S
DeepSeek Coder 1.3B1.35B
Deepseek·16K ctx· coding
369
tok/s · 13% VRAM
›
S
Phi-1 1.3B1.42B
Phi·2K ctx· coding
351
tok/s · 14% VRAM
›
Next step

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your Xbox Series X 6nm GPU.

What this card is worth

Xbox Series X 6nm GPU holds 238 of the models in our catalogue and is, in practice, a Q4_K_M card — the largest it takes is Mistral-Nemo 12.2B at Q4_K_M.

Tokens/sec per $100
—
8B at Q4_K_M, so cards compare like for like
VRAM per $100
—
what memory costs on this card
The biggest it takes
Mistral-Nemo 12.2B12.2B · Q4_K_M9.0 GB42 tok/sest
Mellum2.1 12B-A2.5B Thinking12.15B · Q4_K_M8.4 GB140 tok/sest
Dolly v2 12B12B · Q4_K_S8.7 GB44 tok/sest
StableLM 2 12B12B · Q4_K_M9.0 GB42 tok/sest
Falcon2 11B11B · Q4_K_M8.5 GB45 tok/sest
SOLAR-10.7B10.7B · Q4_K_M8.1 GB46 tok/sest
Falcon3-10B10.3B · Q5_K_M8.8 GB42 tok/sest
GLM-4.1V 9B Thinking10.29B · Q4_K_M8.9 GB48 tok/sest
Against cards of similar memory
M3 Pro (18GB)12 GB$1,5991.5 tok/s per $100
M1 Pro (16GB)11 GB$9992.9 tok/s per $100
M2 Pro (16GB)11 GB$1,2992.2 tok/s per $100
M4 (16GB)11 GB$4994.0 tok/s per $100
Xbox Series X 6nm GPU10 GB——

Neighbours in memory rather than in price: memory decides whether a card can do the job at all, so two cards of the same size at different prices is the comparison you are making. Ordered by memory, then bandwidth — not by the value column, which is worked out from bandwidth and price alone and therefore rewards a cheap card whatever its software stack does to that bandwidth in practice. Read it as one input, not as a ranking.

Device under test

AMD Xbox Series X 6nm GPU — 10 GB VRAM.

Xbox Series X 6nm GPU spec
Brand
AMD
VRAM
10 GB GDDR6
Bandwidth
560 GB/s
FP16 Compute
24.3 TFLOPS
FP32 Compute
12.2 TFLOPS
Stream Processors
3,328
TDP
200 W
Architecture
RDNA 2.0
AI capability
238/ 498 models @ Q4

With 10 GB VRAM and 560 GB/s bandwidth, this GPU handles models up to 12.2B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~64 tok/s.

Rent it — live prices
Checking what the clouds are charging…
Top models for Xbox Series X 6nm GPU
238 fit · showing 20
ModelSizeVRAM Q4tok/sAvg
Mistral-Nemo 12.2B12.2B7.9 GB41129.3
Mellum2.1 12B-A2.5B Thinking12.15B7.9 GB19964.5
Dolly v2 12B12B7.8 GB416.4
StableLM 2 12B12B7.8 GB4121.3
Falcon2 11B11B7.2 GB4533.2
SOLAR-10.7B10.7B7.0 GB4728.2
Falcon3-10B10.3B6.8 GB4838.2
Bamba 9B v29.78B6.5 GB5126.1
Qwen 3.5 9B9.65B6.4 GB5249.2
RecurrentGemma 9B9.63B6.4 GB5235.0
glm-4-9b9.4B6.2 GB5320.5
MiniCPM-o 4.59.37B6.2 GB5381.9
gemma-2-9b9.2B6.1 GB54171.8
Yi 1.5 9B9B6.0 GB5530.3
Yi Coder 9B9B6.0 GB5535.8
K2 Horizon 7B9B6.0 GB5552.4
Ministral 3 8B8.92B5.9 GB5622.6
Ministral 3 8B Reasoning8.92B5.9 GB5669.0
NVIDIA-Nemotron-Nano-9B-v28.9B5.9 GB5644.2
Apertus 1.5 8B8.9B5.9 GB56—